Variability of North American Regulatory Guidance for Heavy Metal Contamination of Residential Soil

Jennings, Aaron A., Petersen, Elijah J., Journal of Environmental Engineering and Science

Abstract: The health implications and remediation obligations of contaminated urban soils are significant issues in many North American cities. Remediation guidance for soils is evolving and shifting from values based on geochemical background approximations to values based on health risks. This analysis examines residential-soil guidance variability among the states and provinces of North America for Cd, Cr, Cu, Ni, Pb, and Zn. For some metals, values vary by as much as 5 orders of magnitude. Nonparametric comparisons are used to illustrate the degree to which current standards differ and to examine the spatial distribution of differences. Parametric statistical analysis is used to examine the significance of variations, and to determine if correlations exist in the structure of guidance values. Hazard index analysis is also used to examine how guidance value differences impact site risk assessments. Results indicate that most of the current variation can be explained as plausible randomness if determining maximum acceptable soil exposures is a random process. Differences between the statistical properties of Cd, Cr, Cu, Ni, and Zn guidance and that of Pb also demonstrate the influence that national leadership can have on stabilizing guidance.

Key words: heavy metals, remediation guidance values, residential soil contamination, risk analysis, hazard index analysis.

Resume : Les repercussions sur la sante et les obligations de restauration des sols urbains contamines sont des questions importantes dans plusieurs villes nord-americaines. Les guides de restauration des sols evoluent et vont de valeurs basees sur les approximations du fond geochimique vers des valeurs basees sur les risques pour la sante. Cette analyse examine la variabilite des guides pour les sols residentiels parmi les etats et les provinces de l'Amerique du Nord pour le Cd, le Cr, le Cu, le Ni, le Pb et le Zn. Pour certains metaux, ces valeurs variaient jusqu'a 5 ordres de grandeur. Les comparaisons non parametriques sont utilisees pour illustrer le niveau de variation des normes actuelles et pour examiner la distribution spatiale des differences. L'analyse statistique des parametres est utilisee pour examiner l'importance des variations et determiner si des correlations peuvent etre etablies dans la structure des valeurs de guide. L'analyse de l'indice de risques est egalement utilisee pour evaluer l'impact des differences de valeurs sur le site. Les resultats montrent que la plus grande partie de la variation actuelle peut etre expliquee par un caractere aleatoire plausible si la determination des expositions maximales aux sols est un procede aleatoire. Les differences entre les proprietes statistiques des guides du Cd, du Cr, du Cu, du Ni et du Zn ainsi que celle du Pb demontrent egalement l'influence que le leadership national peut avoir sur les guides de stabilisation.

Mots cles : metaux lourds, valeurs guides pour la restauration, contamination des sols residentiels, analyse de risques, analyse de l'indice de risques.

[Traduit par la Redaction]

Introduction

The purpose of this study is to examine the variability of current regulatory guidance for heavy metal contamination of residential soil. The analysis will concentrate on current U.S., and Canadian state and provincial environmental agency guidance provided to protect human health.

The motivation for this analysis originated with a 2001 field survey of Cleveland, Ohio area brownfields. The study was designed to identify "old contamination" soils with multicomponent contamination (Jennings et al. 2002a, 2002b). However, because of the extent of contamination identified, (Cd, Cr, Cu, Pb, Ni, Zn maximums of 216, 574, 22500, 15170, 837, and 13400 mg/kg, respectively) the study was expanded to measure contamination at additional brownfields and commons (parks, city gardens, school yards, etc.) distributed throughout the survey area. Information was also developed on the (then) current residential, commercial, and industrial remediation guidelines that could be applied to such sites. The 2001 study identified guidance values from 15 states and provinces and two U.S. EPA regions (Jennings et al. 2002a). Information was also developed on background metal concentrations because "background" was being used as remediation guidance in some states (e.g., NYDEC 1994; OEPA 1999), and because background concentrations add perspective to questions about acceptable contamination levels. This yielded a database of 20 state, province, region, and national estimates of surface-soil background concentrations (Jennings et al. 2002a).

The 2001 survey demonstrated that many Cleveland area commons had heavy metal surface soil concentrations significantly above background levels. This finding led to more extensive commons sampling in 2003, and to an expansion of the regulatory guidelines database. During this 2001-2003 time period, several states revised existing or posted new guidance for residential soil contamination. The 2003 guidance database expanded to 31 states and provinces, 3 EPA regions, and national EPA guidance (Jennings 2003; Petersen 2003).

Discussion of the 2003 guidance database led to the analysis presented here. Several colleagues commented on the degree of variability between states, which was as much as five orders of magnitude for some components. This observation led to questions about the origin of this variation and the impact that variation could have on risk-based remediation assessment.

The analysis presented here addresses the nature, degree, and impact of guidance-value variability for residential soil heavy metal contamination. To accomplish this, the guidance value database was again updated. The 2004/2005 version contains guidance from 52 states and provinces, all of which were confirmed in December of 2004.

Materials and methods

Remediation-guidance values

In this study "remediation guidance values" are defined as numerical values provided by a state or province to help characterize soil contamination during initial stages of site risk assessment. The names used for these values (see Table 1) vary from relatively benign "Preliminary Screening Values" or "Preliminary Remediation Goals" to more assertive "Uniform Remediation Standards", or "Soil Cleanup Levels" designations, but their intents are similar. They are used to identify soils that are sufficiently contaminated to warrant either remediation, or a more advanced site-specific risk analysis. They can serve as remediation standards because, if a soil is remediated to these values, it no longer warrants human-health consideration. In many states, the values serve as explicit Tier 1 "default" remediation standards. The most common origins of guidance values are (i) estimates of background or (ii) estimates of concentrations associated with a specified level of risk. Background guidance values originate from the concept that remediation should return soils to the concentration levels that existed prior to anthropogenic pollution. However, background standards may be unnecessarily conservative, difficult to access (see Background value database section) and awkward for dealing with issues of remediation responsibility. Risk-based guidance values generally originate with a risk assessment that approximates the concentration of a component associated with a specified level of risk. For carcinogenic chemicals, this level is generally a lifetime risk of 1 x [10.sup.5] or 1 x [10.sup.6]. For the metals considered here, the value is most often determined by Hazard index (HI) analysis (see Hazard index analysis section) using a hazard quotient of 1.0, which represents the maximum safe dose for a lifetime of exposure. However, the origins of risk-based numerical guidance values are not always clear. Jennings (2003) explored the origin of Cd, Cr(III), Cr(VI), Cu, Ni, and Zn guidance values from 10 states using their risk analysis formulations and exposure scenario default coefficients. Of the 53 possible comparisons (7 were omitted because guidance was not provided or default data were missing), 24 yielded the state's guidance value ([+ or -]3% to allow for round off), but 29 (55%) did not, and some differences were more than an order of magnitude. Six of these 10 states have since revised their guidance, but not all discrepancies have been resolved. Risk-based guidance values are not always consistent with the specified risk analysis methodology.

Remediation guidance database (2004/05)

The numerical guidance values used in this analysis are summarized in Table 2. Readers are cautioned that the guidance values come from a wide variety of source documents, many of which provide detailed instructions regarding when and how the information should be used. It would not be appropriate to use these numbers without carefully examining the reference documents and regulatory framework from which they have been extracted. Table 2 represents a small amount of the data available in the guidance documents. Values have only been listed for the six metals considered here, and for the soil category most similar to near-surface residential soil (see Table 1 for basis details). Generally, guidance is provided for many organic and inorganic chemicals. For example, the EPA Region III Risk-Based Concentration Table provides guidance for 436 chemicals (USEPAIII 2004). Guidance is also often provided for multiple site categories such as Residential, Commercial, or Industrial land use.

More than one set of guidance values have been listed for California, Massachusetts, New Jersey, and West Virginia. These are alternative or existing and proposed guidance values. In the analysis presented here, only one guidance value set was used for each state. Unused value sets are indicated by italics in Table 2. Guidance values from EPA Regions III, VI, and IX; national EPA; and Environment Canada have also been included, but were only used when they were specified as state or province guidance.

It should be noted that Delaware, Louisiana, Maryland, and Virginia guidance values were increased by a factor of ten (10X) to be comparable with other states because these states base screening on a HI of 0.1. A similar (5X) adjustment was made for Minnesota, North Carolina, and Wisconsin guidance, which base screening on a HI of 0.2. It should also be noted that the values for New York and Missouri are being revised, and for several states (e.g., Missouri, Vermont, California) only draft guidance is available. There have been many recent changes in state guidance values and this trend is likely to continue. In a 1997 survey, Proctor et al. (1997) identified Cr(total), Cr(III), and Cr(VI) residential-soil guidance values for 38 states. Of these 38 value sets, only six (Arizona, Connecticut, Georgia, Maine, Massachusetts, and Oregon) have not changed. Proctor et al. (1997) reported values the states said they were using rather than values published in state guidance documents. Many states were apparently using EPA soil screening levels that have since changed. Most states have now published their own guidance documents, and have modified many of the EPA's recommended values.

Background value database (2004)

Data on background soil composition have been included because these are the origin of some guidance values and because this is a useful baseline for comparison.

One strategy for remediation guidance is to require (suggest, recommend) removal of all pollution above background concentrations. Estimates of background have been used for remediation guidance in several states (OEPA 1999; NYDEC 1994) and seem to influence guidance in several others (SWQCB 2003; GNDR 2003; IDEQ 2004; WDE 2001a, 2001b; AE 1994). Unfortunately, it can be difficult to determine accurate background concentrations because the mineral content of soils can be highly variable and natural metal concentrations have been altered by anthropogenic pollution. It can also be difficult to measure background concentrations of metals, particularly when they are incorporated into the mineral structure of the soil. Variations in extraction methodology add to variability in the background database.

The problems of determining background concentrations can also be compounded by issues of responsibility. Many states consider owners to be responsible for soil contamination that was added as "point source" site activity, but not contamination that was added by regional pollution sources. Some states (e.g., New York) propose remediation to "site background" or "urban background", which combine true background and regional pollution. In addition, background does not correspond to any specific level of risk. "Risk-Based Remediation Guidance" avoids many of the issues of background guidance by seeking remediation to concentrations that yield acceptable risk independently of how the soil was contaminated. Because their origins are unrelated, background-based and risk-based guidance values can differ greatly.

Table 3 presents state and province estimates of background concentrations for the metals considered here. The data are generally for surface soils in rural or agricultural settings, and are averages based on reasonably large numbers of measurements. The Holmgren et al. (1993) data are based on 2771 measurements spread across 32 states (14 to 349 values per state). The Sanders, (2003) New Jersey estimates are from 248 measurements and did not include the 73 measurements of Fields et al. (1993). The Chen et al. (1999) Florida estimates are based on 210 measurements.

The background values of Table 3 are quite variable. Holmgren et al. (1993) applied statistical analysis to determine if differences were significant and found that results from several (often geographically unrelated) states were not significantly different. However, Holmgren et al. (1993), USEPA (2003) and others have used these and similar data to prepare numerically distinct (presumably statistically significant) regional background estimates. For the purposes of the analysis presented here, this was not done because (i) background guidance is being replaced with risk-based guidance so small differences in background concentrations have less significant remediation implications, (ii) significance testing did not consider the variability of replicate state studies or the variability induced by laboratory methodology, and (iii) it is desirable to use background as a baseline for comparing remediation guidance values. This is less convenient if the basis of comparison varies. In this analysis, the "generic" background estimates of Table 3 were used. These values are rationalized, unweighted averages of the state and provincial estimates. Where appropriate, analysis has been added to illustrate differences between results based on this generic background and results based on state-specified background values.

Test case commons

Guidance values may be compared to quantify numerical differences, but this does not gage the importance of differences. Large differences in guidance values for some metals may have very little impact on risk calculations. To provide an additional basis of comparison, actual soil concentrations measured at sample urban sites were selected. Petersen (2003) measured metal concentrations in surface soils at 50 Cleveland, Ohio commons found in the vicinity of brownfields. Thirteen of these sites were selected for comparative analysis. Sites were selected that yielded the minimum, maximum, and near-average concentrations of each metal analyzed (see Table 4). All Cr was assumed to be Cr(III).

Hazard index (HI) analysis

Many risk-based state guidance values for noncarcinogenic chemicals are based on hazard index analysis. This is also a useful basis by which to compare guidance value sets.

Hazard-index values are computed as follows:

[1] HI = [6.summation over (i=1)] [C.sub.i]/[Cs.sub.i] [less than or equal to] (?)1.0

where [C.sub.i] is the measured concentration for the ith metal and [Cs.sub.i] is the maximum concentration of the ith metal that will yield an acceptable lifetime exposure. Generally, the "i" values correspond to Cd, Cr, Cu, Ni, and Zn, and the [Cs.sub.i] values are calibrated for soil ingestion, inhalation, and dermal adsorption (USEPA 1989; Petersen 2003; IDEM 2001; OEPA 2002a, 2002b; ADEC 2002; CDPHE 1997; WDNR 2001a, 2001b).

Each term …

Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Variability of North American Regulatory Guidance for Heavy Metal Contamination of Residential Soil. Contributors: Jennings, Aaron A. - Author, Petersen, Elijah J. - Author. Journal title: Journal of Environmental Engineering and Science. Volume: 5. Issue: 6 Publication date: November 2006. Page number: 485+. © 2008 NRC Research Press. COPYRIGHT 2006 Gale Group.

This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.